(S)-3-Amino-1-benzylpyrrolidine and (S)-3-[(S)-2-{tert-butoxycarbonylamino}-propionylamino] pyrrolidine are key intermediates required for the synthesis of several quinolone antibacterial agents disclosed, for example, in U.S. Pat. Nos. 4,851,418 and 4,916,141. Known processes for the synthesis of optically active 3-amino-1-substituted-pyrrolidines include the conversion of 1-benzyl-3-pyrroline to (S)-3-amino-1-benzylpyrrolidine in four steps in 60% overall yield with an enantiomeric excess of 84% (H. C. Brown, J. V. N. Vara Prasad, A. K. Gupta, J. Org. Chem., 51, 4296 (1986); T. Rosen et al, J. Med. Chem., 31, 1586 (1988); D. T. W. Chu, T. J. Rosen, European Patent Application EP 331,960). This process uses the relatively expensive hydroboration reagent, diisopinocampheylborane. In a second process trans-4-hydroxy-L-proline is converted into (S)-3-amino-1-(benzyloxycarbonyl)-pyrrolidine in five steps in 77% overall yield with an enantiomeric excess greater than 99%, U.S. Pat. No. 4,851,418; however, the trans-4-hydroxy-L-proline starting material is expensive. Racemic 3-amino- 1-benzylpyrrolidine has been resolved in low yield by fractional crystallization of the salt with L-tartaric acid (Tokyo Ksei Kogyo, Japanese Patent J02218-664A (1989). Chiral butane derivatives with leaving groups (chloro or methanesulfonyloxy) at positions 1, 2, and 4 were treated with primary amines to give chiral 3-(substituted amino)-1-substituted pyrrolidines (Tokyo Kasei Kogyo, Japanese Patent J91020-391-B (1987); European Patent Publication 443498)). Finally the relatively inexpensive L-aspartic acid was converted to (S)-3-amino-1-phenylpyrrolidine via 3-[N-(tert-butoxycarbonyl)amino]-1-phenylpyrrolidine-2,5-dione, although in low overall yield (D. T. Witiak, et al, J. Med. Chem., 14, 24 (1971).
Due to its low cost, L-aspartic acid appears to be the most attractive starting material for the preparation of (S)-3-amino-1-substituted-pyrrolidines. The literature method from L-aspartate (D. T. Witiak, loc. cit.) involves treatment of N-boc-L-aspartate with acetic anhydride to give (S)-3-(tert-butoxycarbonylamino)succinic anhydride followed by addition of a primary amine and ring closure with heat to give the succinimide, (S)-3-(tert-butoxycarbonylamino)-1-substituted pyrrolidine-2,5-dione. The latter is then reduced and deblocked to give the 1-substituted-3-aminopyrrolidine in low overall yield. From our experience with this method some racemization takes place when the blocking group is p-toluenesulfonyl(tosyl) and the primary amine is benzylamine.
We have now found an alternate, high yield method for converting L-aspartic acid to (S)-3-amino-1-substituted pyrrolidines which involves the synthesis of (S)-2-(2'methanesulfonyloxyethyl)-1-(p-toluenesulfonyl)aziridine and its conversion to (S)-1-benzyl-3-(p-toluenesulfonylamino)pyrrolidine as the key steps. The method may be extended to the synthesis of other optically active 3-amino pyrrolidines with different substituents (R') at the 1 position.